Article Text


Epidemiology oral session 8: STIs and HIV in female sex workers
O1-S08.02 Who are the women at risk of HIV infection in rural Zimbabwe and how many are there? Insights into their characteristics, locations, and behaviours
  1. J Elmes1,
  2. K Nhongo2,
  3. T Hallett1,
  4. P White1,
  5. R Mutsindiri2,
  6. G Garnett1,
  7. C Nyamukapa2,
  8. S Gregson1
  1. 1Imperial College London, London, UK
  2. 2Biomedical Research and Training Institute, Zimbabwe


Background In Zimbabwe's generalised epidemic, the relative importance of different risk groups to ongoing HIV transmission is unclear. We estimated population sizes and HIV prevalence of women involved in different levels of sex work activity in two socioeconomic locations (rural areas and commercial centres) in east Zimbabwe.

Methods The baseline survey for a cohort study of 650 women at high risk of HIV infection (WR) was conducted in March 2010, using snowball and location-based methods to recruit women with any form of sexual exchange motivated by material reward including cash-per-act and informal credit-based arrangements. We restricted our analysis to a subset of WR who report receiving money for sex, excluding non-monetary transactions. WR were assigned to high, medium or low activity levels based on, among other factors, the frequency they report receiving money for sex and numbers of partners. CIs are based on Agresti-Coull 95% estimation and p values are for χ2 estimation unless otherwise stated.

Results WR were more common in commercial centres than in rural areas 11.1% (9.9% to 12.4%) vs 5.0% (4.3% to 5.8%). The pattern of sex work activity differed between locations (p=0.049) and mean number of sexual partners, over 2 weeks, in high, medium and low activity WR were 4.1, 1.9 and 1.3 respectively (ANOVA test for trend: p<0.001). High activity WR comprised 2.4% (1.9% to 3.1%) of women in commercial centres vs 0.55% (0.34% to 0.89%) in rural; medium level, 7.1% (6.2% to 8.2%) vs 3.9% (3.3% to 4.7%); and low level, 1.56% (1.13% to 2.13%) vs 0.50% (0.30% to 0.82%),Abstract O1-S08.02 figure 1. Overall, HIV prevalence in WR was higher in commercial centres than in rural areas: 46.5% (95% CI: 38.9% to 54.3%) vs 37.8% (28.4% to 48.1%) (p>0.1). Mean HIV prevalence among high level WR was consistent across socioeconomic strata: 50% (range: 35.8%–64.2%). Prevalence among medium level WR tended to be higher in business centres: 47.5% (38.1% to 57.2%) vs 35.2% (25.1% to 46.8%) in rural areas. In contrast, HIV prevalence in low level WR was higher in rural areas: 44.4% (18.8% to 73.4%) vs 36.4% (19.6% to 57.1%).

Abstract O1-S08.02 Figure 1

Relative size and HIV prevalence of the WR population in two socioeconomic locations in east Zimbabwe. Circle sizes are in proportion to the relative size of WR activity level; the inner ring represents low activity, the central ring high activity and the outer ring medium activity WR.

Conclusions Sex worker populations may be larger than previously measured in general population surveys in rural Zimbabwe. Given that HIV prevalence among WR is threefold higher than in the general population and the underestimated size of the sex worker population, sex work may be an increasingly important driver of HIV transmission in declining generalised rural African epidemics.

Statistics from

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.